[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Yes, they are special, but...
Date: Mon, 12 Jun 89 10:10 EDT
From: Scott McKay <SWM@SAPSUCKER.SCRC.Symbolics.COM>
Date: Sat, 10 Jun 89 03:25:17 CDT
More nitty-gritty: you can't run multiple processes in separate address
spaces on the LispM. I think this is actually a fairly serious
limitation for some applications. (Besides the fact that it means the
CPU can do little more than twiddle its thumbs during page faults.) It
certainly keeps the machine from being effective in multi-user
applications. (Again, don't get me wrong. I don't fundamentally
disagree with the "one user, one CPU" philosophy. I just think of it as
a first approximation, not a rigid rule.)
"Multiple processes in a single address space" is a function of Genera,
not the Lisp machine hardware.
Are you saying the hardware can in fact run multiple address spaces?
I have no argument or fixed belief to the contrary, but I've certainly
never heard anyone say that.
"Multiple processes in a single address
space" does not imply that "the CPU can do little more than twiddle its
thumbs during page faults". In fact, we have machines here which have
software which allows other processes to run during a page fault.
However, metering the performance effects of that change is very
difficult. We believe that, under many common conditions, scheduling
across page faults improves performance, but there are other conditions
under which it degrades performance.
Seems to me this would depend on what you mean by "performance". I
can imagine an occasional degradation in response time, but have
trouble imagining that throughput would ever suffer. This is the same
tradeoff timesharing systems have always made, of course.
But the point I was really making is not that timesharing improves
performance, but that it improves *functionality*. It's essentially
impossible right now to use a LispM over a dialup; not that I would
want to do so very often, or for very long at a time, but there are
those occasions when it sure would be handy. I know, I know, multiple
address spaces and scheduling across page faults are not absolute
requirements for that by any means. But they do make it work better.
And anyhow, as you acknowledge, they are the kind of thing the
workstation culture expects.